A ceramide-1-phosphate analogue, PCERA-1, simultaneously suppresses tumour necrosis factor-α and induces interleukin-10 production in activated macrophages

Reagents and cell culture

LPS (Escherichia coli serotype 055:B5), imiquimod, peptidoglycan, isobutylmethylxanthine (IBMX), LPA, polymyxin-B, phenylmethylsulphonyl fluoride (PMSF) and dimethyl sulphoxide (DMSO) were purchased from Sigma-Aldrich (St Louis, MO). Trypsin, l-glutamine, penicillin and streptomycin were purchased from Biological Industries (Beit Haemek, Israel). Dulbecco’s modified Eagle’s minimal essential medium (DMEM) and fetal bovine serum (FBS) were purchased from Gibco (Carlsbad, CA). Bovine serum albumin (BSA) was purchased from Amresco (Solon, OH). C8-ceramide-1-phosphate (C8-C1P) and sphingosine-1-phosphate (S1P) were purchased from Avanti Polar lipids (Alabaster, AL), dissolved in ethanol and then diluted in culture medium containing 4% fatty acid-free BSA. A neutralizing monoclonal anti-mouse IL-10 antibody, rat immunoglobulin G (IgG) isotype control, recombinant mouse TNF-α (rTNF-α) and enzyme-linked immunosorbent assay (ELISA) reagent sets for TNF-α and IL-10 were purchased from R&D Systems (Minneapolis, MN). The cAMP enzyme immunoassay (EIA) kit was purchased from Cayman Chemicals (Ann Arbor, MI). Antibodies against doubly phosphorylated mitogen-activated protein (MAP) kinases, general MAP kinases, α-tubulin and proliferating cell nuclear antigen (PCNA) were obtained from Sigma, while the antibody against the p65 subunit of NF-κB was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Infrared dye-labelled secondary antibodies and blocking buffer were obtained from Li-Cor Biosciences (Lincoln, NE). Nitrocellulose membranes were purchased from Sartorius (Gollingen, Germany) and Immobilon-FL polyvinylidene fluoride (PVDF) membranes were purchased from Millipore (Billerica, MA). Complete protease inhibitor mixture was purchased from Roche (Mannheim, Germany). NAP-5 columns were purchased from GE Healthcare (Little Chalfont, UK). PCERA-1 and its non-phosphorylated analogue, CERA-1, were synthesized according to published procedures.^21,29 PCERA-1 was dissolved in phosphate-buffered saline (PBS) and freshly diluted in culture medium, while CERA-1 was initially dissolved in ethanol and then diluted in culture medium containing 4% fatty acid-free BSA. Mouse RAW264.7 macrophage cells, obtained from the American Type Culture Collection (ATCC, Rockville, MD), were grown to 80–90% confluence in DMEM medium supplemented with 2 mm l-glutamine, 100 U/ml penicillin and 100 μg/ml streptomycin (hereafter culture medium) and 10% FBS, at 37° in a humidified incubator with 5% CO₂.

Lysis buffers

Buffer A contained 1% Triton X-100, 50 mm Tris (pH 8·0), 100 mm NaCl, 50 mmβ-glycerophosphate, 40 mm NaF, 1 mm sodium orthovanadate, 1 mm ethylenediaminetetraacetic acid (EDTA), 1 mm ethyleneglycoltetraacetic acid (EGTA) and 30% glycerol. Buffer B contained 10 mm HEPES, pH 7·9, 10 mm KCl, 0·25 μm dithiothreitol (DTT) and 0·1 mm EDTA. Buffer C contained 1% NP-40, 0·025% sodium dodecyl sulphate (SDS), 1% deoxycholic acid, 50 mm Tris, pH 8·0, 0·4 m NaCl and 0·1 mm EDTA. In addition, all lysis buffers contained ‘complete’ protease inhibitor mixture, diluted according to the manufacturer’s instructions, and supplemented with 1 mm PMSF.

Determination of free and protein-bound (PCERA-1)

PCERA-1 (10 μm) was incubated for 2 hr at 37° in culture medium, with or without serum (5% FBS). Free PCERA-1 and protein-bound PCERA-1 were completely separated on a NAP-5 exclusion chromatography column. A collection efficiency of 56% for the low-molecular-weight fraction was determined by mass spectrometry (MS), using PCERA-1 (in the absence of serum proteins) as a probe. A collection efficiency of 88% for the high-molecular-weight fraction was determined by a modified Bradford protein assay (details below), with serum proteins as a probe. For release and quantification of protein-bound PCERA-1, the high-molecular-weight fraction was diluted 1 : 4 with methanol at −20°, vortexed for 1 min, and concentrated back to the original volume using a speed-vac.

The concentration of PCERA-1 in the two fractions was measured by liquid chromatography mass spectrometry (LC-MS). Briefly, the PCERA-1 sample (5 μl) was applied to a SB-C18 column (Agilent Zorbax, 2·1 mm × 75 mm; Agilent Technologies Inc., Santa Clara, CA) on an Agilent 1100 LC system. Solvents A and B were 0·1% formic acid in H₂O or in acetonitrile, respectively, and the flow rate was 0·25 ml/min. Isocratic elution with 90% solvent A for 3 min was followed by a linear gradient ramping to 100% solvent B at 15 min, with the column effluent being sent to an Agilent 6410 triple quadrupole mass spectrometer. PCERA-1 was quantified by monitoring the signature m/z transition state 388 → 290 and using 388 → 164 as a qualifying transition. Analysis was carried out in positive ion mode with a capillary voltage of 4000 V, a fragmentor voltage of 115 V and a dwell time of 250 milliseconds, and a collision energy of 15 V. A calibration curve of PCERA-1 standards (2·5–50 pmol) was obtained before and after all samples, to correct for signal drift. Duplicate analysis of each sample was used to determine an average and standard deviation, and the whole experiment was repeated twice.

Macrophage activation assay

RAW264.7 macrophages were maintained for 48 hr prior to the experiment in 96-well plates, at 2 × 10⁵ cells/well, in culture medium supplemented with 5% FBS, at 37° in a humidified incubator with 5% CO₂. The cells were stimulated with LPS (10–100 ng/ml) at 37° for 2 hr in the presence or absence of PCERA-1 (10 μm). TNF-α and IL-10 secretion to the medium was measured by ELISA.

Cytokine and cAMP measurements

Measurements of TNF-α and IL-10 levels in supernatants of RAW264.7 cells were performed with commercially available ELISA reagent sets, according to the manufacturer’s instructions, using a microplate reader (Bio-Tek, Winooski, VT). The cytokines were undetectable (< 20 pg/ml) in the absence of LPS. The samples were stored at −80° until use. All experiments were repeated as least three times. Measurements of intracellular cAMP levels in RAW264.7 cells were performed with a commercially available EIA kit, according to the manufacturer’s instructions.

Viability assay

RAW264.7 macrophages were seeded in a six-well plate at 8 × 10⁵ cells/well and cultured for 48 hr as above for the activation assay. The cells were treated with LPS (100 ng/ml) and/or PCERA-1 (10 μm) for 2 hr at 37°, harvested with cold PBS, centrifuged (1000 g for 3 min), and re-suspended in PBS at 1·6 × 10⁶ cells/ml. Viability was measured by FACS (Easy Cyte Mini System; Guava, Hayward, CA) using a Guava Viacount reagent kit, according to the manufacturer’s instructions. Analysis was performed using the Guava cyto soft software.

MAP kinase phosphorylation assay

RAW264.7 macrophages were maintained for 24 hr prior to the experiment in 12-well plates, at 8 × 10⁵ cells/well, in culture medium supplemented with 0·1% FBS. The cells were stimulated with LPS (100 ng/ml) at 37° for 30 min in the presence or absence of PCERA-1 (10 μm). The cells were washed twice with cold PBS and lysed for 1 hr at 4° with buffer A. Cell extracts were centrifuged (14 000 g, 15 min at 4°) and the supernatants were stored at −80°.

NF-κB activation assay

RAW264.7 macrophages (3 × 10⁶ cells per T-25 flask) were grown in culture medium supplemented with 10% FBS for 48 hr, up to a confluence of 90%. The cells were stimulated with LPS (1 μg/ml) at 37° for 1 hr in the presence or absence of PCERA-1 (0·1–10 μm). The cells were harvested with trypsin, centrifuged (1000 g for 3 min), and washed twice with cold PBS. Separation of the cytosolic and nuclear fractions was accomplished by a modification of a procedure of Zilberman et al.³⁰ Briefly, cell pellets were re-suspended in lysis buffer B, and allowed to swell on ice for 15 min. Then, 2·5% NP-40 was added, and the samples were centrifuged at 900 g for 5 min. The cytosolic fraction was collected and stored at −80° until use, while the nuclear pellet was washed again with buffer B, re-suspended in buffer C and homogenized using a Dounce Wheaton homogenizer (Wheaton Science, Millville, NJ) and tight pestle. The nuclear extracts were kept on ice for 1 hr with intermittent vortexing prior to centrifugation at 20 000 g for 10 min. The supernatant (nuclear fraction) was stored at −80° until use.

Western blotting

Cell extracts (30 μg protein) were boiled for 5 min in sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) buffer and subjected to 10% SDS-PAGE, and proteins were transferred to a nitrocellulose (NF-κB assay) or Immobilon-FL PVDF (MAP kinase assay) membrane. Two-colour imaging and quantitative analysis of western blots were performed using the Odyssey infrared imaging system (Li-Cor Biosciences), according to the manufacturer’s instructions.

Protein determination

Protein was determined by a modification of the Bradford procedure, which yields linear results, increased sensitivity, and reduced detergent interference, as previously described by Zor and Selinger.³¹ BSA served as a standard.

Quantitative reverse transcription–polymerase chain reaction (RT-PCR)

The mRNA levels of TNF-α and IL-10 in RAW264.7 cells were quantified by real-time PCR. The cells were seeded in a six-well culture plate at 1 × 10⁶ cells/well and cultured for 48 hr in culture medium supplemented with 5% FBS. The cells were then treated with LPS (100 ng/ml) with or without PCERA-1 (10 μm) for 1 hr at 37°. Total RNA was isolated using the MasterPure RNA purification kit (Epicentre Biotechnologies, Madison, WI), and 1 μg of RNA from each sample was reverse transcribed into cDNA using the Verso cDNA synthesis kit (Thermo Scientific, Waltham, MA). Quantification of cDNA (50 ng total) was preformed on the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA), using TaqMan probes, purchased from Qiagen (Valencia, CA) for TNF-α measurement, and from primerDesign (Southampton, UK) for IL-10 and actin measurements.

Statistical analysis

All the data were analysed using Student’s t-test wherever applicable. In all cases, differences of P < 0·05 were considered to be significant. Effective concentration (EC₅₀) values were calculated by non-linear regression curve fitting, using the sigmaplot software (Systat Software Inc., San Jose, CA).

In vitro activity of PCERA-1 in RAW264.7 macrophage cells

To assess the in vitro effect of PCERA-1 on LPS-stimulated TNF-α secretion, mouse RAW264.7 macrophages were incubated with various concentrations of PCERA-1 and stimulated with LPS for 2 hr. Figure 2(a) shows that PCERA-1 suppressed production of the pro-inflammatory cytokine TNF-α with an EC₅₀ of 100 ± 20 nm (mean ± SD of 3 independent experiments), and that maximum inhibition was observed at a concentration of approximately 1·0 μm. Interestingly, PCERA-1 increased in vitro production of the anti-inflammatory cytokine IL-10, with an identical EC₅₀ of 100 ± 20 nm (Fig. 2a). As PCERA-1 is a phospholipid-like molecule, it is expected to bind to serum albumin present in the culture medium. Such binding decreases the actual concentration of PCERA-1 which is free and available for interaction with the receptor. We thus measured the concentration of free PCERA-1 in the serum-supplemented culture medium, using size exclusion chromatography (separating protein-bound and protein-free PCERA-1) followed by quantification using LC-MS analysis. The distribution of PCERA-1 (10 μm total) between the protein-bound and protein-free fractions was 3888 ± 605 and 628 ± 37 pmol, respectively, while total PCERA-1 (in the absence of serum) was 5000 ± 953 pmol (mean ± SD). Thus, roughly 87 ± 3% of PCERA-1 (at 10 μm) is bound to serum proteins, indicating that the actual EC₅₀ for free PCERA-1, available for interaction with its receptor, is 13 ± 2 nm (calculated as 13% of 100 nm). This calculation represents a maximal estimation because at the total concentration of 100 nm, which yields a 50% effect, a lower fraction of PCERA-1 may be protein-free than at 10 μm. Yet, we cannot rule out the possibility that protein-bound PCERA-1 is also active.

Attesting to specificity, the dephosphorylated form of PCERA-1 (CERA-1) had no significant effect on the production of TNF-α or IL-10 (Fig. 2b). Furthermore, no significant effect on cytokine production was observed for C8-ceramide-1-phosphate (C8-C1P), a phospholipid that has close structural similarity to PCERA-1 (1, 2), or for the native extracellular phospholipid mediators S1P and LPA, which are also related in structure to PCERA-1 (Fig. 2b). Finally, neither PCERA-1 nor LPS affected macrophage viability (75 ± 5% for untreated or treated cells).

We examined whether the continuous presence of PCERA-1 was necessary for the suppression of LPS-induced TNF-α production. Preincubation of the macrophages with PCERA-1 followed by its removal prior to LPS addition resulted in a loss of PCERA-1 activity (Fig. 3a). Similarly, prostaglandin E₂ (PGE₂), an agonist of a cell-surface receptor, suppressed TNF-α production only when co-incubated with LPS. In contrast, dexamethasone, an agonist of a nuclear receptor, efficiently suppressed TNF-α production even when the LPS stimulus was applied after its preincubation and removal from the cell culture medium (Fig. 3a).

Significantly, preincubation of the macrophages with PCERA-1, prior to LPS addition, was not required for its activity, as evident from the identical activities of PCERA-1 when added 30 min before LPS (without removal) or co-added with LPS (Fig. 3b). However, the ability of PCERA-1 to suppress prior LPS-induced TNF-α production was strictly time-dependent. Full inhibitory activity was demonstrated when PCERA-1 was added no later then 10 min after LPS, while a delay of 1 hr resulted in a major loss of its activity, although LPS incubation time was increased from 2 to 3 hr (Fig. 3b). The ability of PCERA-1 to amplify LPS-induced IL-10 production was similarly time-dependent (data not shown). Taken together, our results demonstrate that PCERA-1 can specifically exert a robust anti-inflammatory effect on LPS-stimulated macrophages in culture, in an extracellular time-dependent manner.

TNF-α suppression and IL-10 induction are independent

As the modulations of TNF-α and IL-10 production by PCERA-1 had identical concentration dependences (Fig. 2a) and were found to be on similar time scales (data not shown), we hypothesized that PCERA-1 binds a single receptor that reciprocally regulates production of TNF-α and IL-10. In order to experimentally determine whether indeed both TNF-α suppression and IL-10 induction are primary effects of PCERA-1 or whether one of them is a secondary effect of the other, we measured cytokine production by LPS-stimulated RAW264.7 cells treated with or without PCERA-1 in the presence or absence of a neutralizing anti-IL-10 antibody or a recombinant TNF-α protein (rTNF-α). The anti-IL-10 antibody was added to eliminate the possible TNF-α suppression mediated by PCERA-1-induced IL-10, while rTNF-α was added to eliminate the possible effect of the PCERA-1-suppressed TNF-α level on IL-10 production. We found that PCERA-1-mediated suppression of TNF-α production was not affected by the anti-IL-10 antibody (Fig. 4a), although the antibody totally sequestered the secreted IL-10 (data not shown). An IgG isotype-matched control antibody also did not significantly affect the level of either TNF-α or IL-10. This result indicated that TNF-α suppression is not mediated by PCERA-1-induced IL-10 in these cells. Similarly, addition of exogenous rTNF-α to PCERA-1-treated cells, up to the level of LPS-induced TNF-α obtained in the absence of PCERA-1, did not change IL-10 secretion (Fig. 4b). We therefore conclude that IL-10 induction and TNF-α suppression are independent effects of PCERA-1 in activated macrophages.

PCERA-1-mediated modulation of cytokine production by agonists of TLR2, TLR4 and TLR7

To further establish the general anti-inflammatory character of PCERA-1 activity, we incubated RAW264.7 cells with or without PCERA-1 in the presence of either LPS (a TLR4 agonist), peptidoglycan (a TLR2 agonist) or imiquimod (a TLR7 agonist). As shown in Fig. 4(c), TNF-α suppression by PCERA-1 was general and independent of the specific TLR activated. Additionally, IL-10 production was markedly up-regulated by PCERA-1, in the presence of any of the TLR agonists (Fig. 4c). Induction of IL-10 by PCERA-1 in the absence of a TLR agonist was as low as induction by the TLR agonist alone, indicating a synergistic IL-10 production (data not shown). To rule out a possible contamination of LPS in the peptidoglycan and imiquimod samples, we added polymyxin-B, a specific LPS blocker, and observed diminished cytokine production in the presence of LPS but not in the presence of the TLR2 or TLR7 agonist (data not shown).

Intracellular signalling of PCERA-1

To provide further insight into the molecular mechanism underlying the PCERA-1-dependent reciprocal regulation of TNF-α and IL-10 release, the mRNA levels of these cytokines in RAW264.7 cells were quantified by real-time RT-PCR. Figure 5 shows that PCERA-1 reduced the LPS-induced TNF-α mRNA level by 33%, while it elevated the IL-10 mRNA level 2·7-fold over the LPS-induced level. Actin served as a control for the quantity of total mRNA analysed. These results suggest that PCERA-1 is involved in regulation of TNF-α and IL-10 gene expression in RAW264.7 cells.

LPS-induced production of TNF-α depends primarily on the transcription factor NF-κB,³² and on the mitogen-activated protein (MAP) kinases p38, extracellular signal-regulated kinases (ERK) 1/2 and c-jun N-terminal kinase (JNK).³³ We therefore sought to examine the effect of PCERA-1 on LPS-induced activation of these key mediators. Figure 6(a) shows that PCERA-1 was unable to inhibit the LPS-induced nuclear translocation, and hence activation, of the p65 subunit of NF-κB. Figure 6(b) shows that PCERA-1 was unable to significantly affect the LPS-induced phosphorylation, and hence activation, of any of the three examined MAP kinases. It should be noted that PCERA-1 did not block activation of these signalling pathways, when measured either at the peak of activation (Fig. 6; 1 hr for NF-κB and 0·5 hr for MAP kinases) or at earlier time-points (data not shown). The relatively low basal activation of NF-κB, and even lower basal activation of MAP kinases, may be attributable to serum factors or growth conditions in general.

Finally, as cAMP is known to mediate TNF-α suppression³⁴ and IL-10 induction⁷ through a multitude of endogenous extracellular mediators (i.e. PGE₂), we sought to determine whether PCERA-1 affects the cAMP level. Figure 7 shows that PCERA-1 indeed elevated the intracellular cAMP level fivefold relative to the vehicle. The observed cAMP increase was LPS-independent and comparable to that induced by PGE₂ (data not shown). The structurally related phospholipid C8-C1P (Fig. 1) had no significant effect on cAMP level (Fig. 7). Taken together, our results suggest that PCERA-1 specifically increases intracellular cAMP to a functionally significant level.